Thyroid Function and Obesity in US Pediatric Population

Presentation Number: THR-543
Date of Presentation: March 5th, 2015

Tara Kaushal*1, Morri Markowitz2, Jinny Cai3 and Ping Zhou4
1The Children's Hospital at Montefiore, Bronx, 2Albert Einstein College of Medicine, Bronx, NY, 3University of Southern California, 4Children's Hospital at Montefiore, Albert Einstein College of Medicine, Bronx, NY



 Numerous studies have pointed out an association between obesity and abnormalities in thyroid function tests (TFT).  The results are, however, inconsistent.  Thyroid hormone and thyroid-stimulating hormone (TSH) concentrations have been described as normal, elevated or reduced in adult obese individuals in different studies.  In obese children, the most common thyroid abnormality is hyperthyrotropinemia.  Population-based studies for pediatric subjects are lacking.  The U.S. National Health and Nutrition Examination Survey (NHANES) conducted in 2001 to 2002 provided a unique opportunity for such a study, because thyroid hormone and thyroid antibody measures were sampled in the pediatric population.


Using a large, national data set (NHANES 2001-2001), to examine the relationship of thyroid function and weight in a pediatric population, and compare the results with the previous studies.

Method Design, Data Management, and Analysis

From NHANES 2001-2002, data on anthropomorphic and thyroid related measures (TSH, T4, FT4, T3, FT3, anti-thyroglubin antibody, and anti-thyroid peroxidase antibody) were extracted from the files DEMO_B, L06VID_B, and SSNH4THY.  Only pediatric subjects between ages of 12 and 20 with negative thyroid antibodies were selected for the study.  BMI Z-scores were calculated by using the CDC BMI percentile for specific age and gender.

A Pearson correlation matrix was calculated to examine the relationship between BMI Z-score and thyroid hormone measures.  We also divided the group into 4 BMI subgroups: lean, normal, overweight, and obese (0.0%-4.9%, 5.0%-84.9%, 85.0%-94.9%, and =>95.0% respectively). The means of thyroid hormone measures are compared among the groups by using t-tests.

All analyses were performed in a SAS statistical package (Version 9.3, SAS Institute, Cary, NC).


 Data from 611 subjects were included. BMI Z-score showed significant positive correlations with T3 (r=0.17, p<=0.0001), TSH (r=0.11, p=0.008), and FT3 (r=0.10, p=0.02), but no significant correlation with T4 and FT4 .

Compared with the normal group, the means of T3 (129 vs. 141), FT3 (3.6 vs. 3.8), TSH (1.48 vs. 1.79), and T4 (8.1 vs. 8.5) in our obese group are significantly higher with p-values of  <0.01, 0.004, 0.02, and 0.03 respectively.

No statistically significant differences were found in comparisons between the normal group and either the lean or overweight groups, except that the mean FT4 level is slightly lower in the overweight group.  However, this difference was not found in the comparison with the obese group.   

Overall, thyroid functions of our subject population fall in normal reference range which are 5.4 -12.8 mcg/dl for T4 and 0.47-5.01 mIU/L for TSH.


 Our analyses suggest that TSH and thyroid hormone levels are higher in the obese pediatric population.  However, overall thyroid functions remain in normal ranges.


Nothing to Disclose: TK, MM, JC, PZ